The proposed research is a longitudinal epidemiological study of the influence of race and social factors on functional disability in community-residing elderly persons. It will examine whether: a) Blacks are at higher risk of decline in functioning than Whites, b) whether socioeconomic factors, such as income and education, and prevalent and incident chronic illness account for the differences, and c) whether the risk of disability among Blacks changes after the age of 75 or 85. In addition, the research will examine social factors in relation to disability, in particular social network and social support, hypothesizing that elderly persons with fewer social ties and less support are at higher risk of functional disability. Also, the combined effect of race and social network and support on disability will be tested, addressing whether the lack of network ties and social support increase the risk of functional decline more so among elderly Blacks than Whites. Data for the research come from two multiwave panel studies of representative community-samples of adults aged 65 and over in New Haven and North Carolina, which are two sites of the N.I.A. funded EPESE Project. Each study consisted of a baseline interview and annual reinterviews, with repeated comprehensive assessments of functional disability and major chronic illnesses. Katz' Activities of Daily Living index, which is a measure of self-care activities (e.g. grooming and eating), win serve as main outcome variable. Information on other aspects of functioning is available as well, such as gross mobility (e.g. walking 1/2 mile), physical performances (e.g. kneeling, extending arms above. shoulder), and daily activities other than self-care (e.g. cooking, shopping). This information will be used in some analyses to study course of disability in more detail. Two outcomes will be identified: onset and course of disability. Onset will be defined as development of nonreversing disability in self-care functions among those free of disability at baseline. Course of disability will be defined on the basis of changes in disability during follow-up for the entire cohorts, and will not be limited to nonreversing disability. The research will employ standard epidemiologic analytic techniques, such as logistic regression and proportional hazards models. For some analyses, grade-of-membership analysis will be employed, which is a statistical technique that is especially appropriate for the classification and analysis of health characteristics in heterogeneous populations.